An official website of the United States government
Here's how you know
Official websites use .gov
A
.gov website belongs to an official
government organization in the United States.
Secure .gov websites use HTTPS
A lock (
) or https:// means you've safely
connected to the .gov website. Share sensitive
information only on official, secure websites.
As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with,
the contents by NLM or the National Institutes of Health.
Learn more:
PMC Disclaimer
|
PMC Copyright Notice
Exploring the relationship among adverse events is important because those that arise from a common mechanism are amenable to a common intervention, which can improve symptom management, quality of life, and treatment adherence. To date, symptom cluster studies have used patient-reported data, which are not always available in clinical trials. In this study, we proposed using the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) to identify adverse event clusters because the CTCAE are collected as standard practice and can therefore be used when patient-reported outcomes are unavailable. Hence, is it feasible to identify severe adverse events clusters from data captured using the CTCAE in clinical trials?
SUMMARY ANSWER:
Six severe adverse events clusters were identified in patients with advanced prostate cancer. Identifying adverse events clusters using CTCAE data from clinical trials is feasible.
METHODS:
A variable-based hierarchical cluster analysis was conducted using the CTCAE data captured from 323 patients who experienced at least one grade 3 or higher adverse event in an advanced prostate cancer randomized clinical trial conducted by SWOG (S9916).
BIAS, CONFOUNDING FACTOR(S), DRAWBACKS:
The difficulty of using adverse event data from clinical trials is that often not all adverse events are recorded. In our study, only the highest severity grade for each adverse event type was recorded, and only grade 3 or higher adverse events were captured reliably. Moreover, in contrast to previous publications on symptom cluster that used patient-reported outcomes, the CTCAE is clinician reported and may not accurately reflect the presence of patient symptoms and the severity of these symptoms.
REAL-LIFE IMPLICATIONS:
Capturing adverse events using the CTCAE, which is standard practice in all clinical trials, can be used to understand the relationships among adverse events and to identify adverse events clusters when patient-reported outcomes are unavailable.
Dendrogram for the cluster analysis. PRBC, packed RBC; w/o, without. Numbers in the figure are Spearman rank correlation, and Ps are from the Mantel-Haenszel χ2 test.
Exploring the relationships among adverse events is important because those that arise from a common mechanism are amenable to a common intervention, which can improve symptom management, quality of life, and treatment adherence. To date, symptom cluster studies have used patient-reported data, which are not always available in clinical trials. In this study, we proposed using the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) to identify adverse event clusters because the CTCAE data are collected as standard practice and can therefore be used when patient-reported outcomes are unavailable.
Methods:
The CTCAE data from a randomized clinical trial conducted by SWOG that compared docetaxel plus estramustine versus mitoxantrone plus predinsone in patients with advanced prostate cancer were used to identify severe adverse event clusters. A variable-based hierarchical cluster analysis was conducted using the CTCAE for the 323 patients who experienced at least one grade 3 or higher adverse event.
Results:
A total of 109 adverse event types were captured using the CTCAE. Four clusters had moderate associations: nausea, vomiting, and anorexia (n = 35, r = 0.45); joint/bone (myalgia, arthralgia, and arthritis) and muscle weakness (n = 26, r = 0.29); anemia and transfusion (n = 20, r = 0.38); and neutrophils/granulocytes, febrile neutropenia, and leukocytes/lymphopenia (n = 114, r = 0.29). Two clusters had weak associations: fatigue/malaise/lethargy and dehydration (n = 66, r = 0.12); and constipation, infection without neutropenia, and abdominal pain/cramping (n = 35, r = 0.13).
Conclusion:
Several severe adverse event clusters were identified in patients with advanced prostate cancer. Identifying adverse event clusters using CTCAE data from clinical trials is feasible.
INTRODUCTION
With an increasing number of cancer therapies being prescribed for an extended amount of time (months and even years), understanding the adverse events resulting from these therapies and the symptoms experienced by patients with cancer is critical. Patients with advanced cancer often experience multiple symptoms caused by treatments and therapies. Traditionally, in cancer research, such as prostate research, symptom management has focused on individual symptoms such as incontinence, erectile dysfunction, fatigue, hot flashes, pain, and depression,1-4 but the interrelation among such symptoms is important and has not been explored thoroughly.5,6 This need has been addressed by recent publications in the area of symptom cluster research; a symptom cluster is typically defined as two or more related symptoms that may or may not come from a common source but that occur simultaneously.7,8 Recognizing these symptom clusters and exploring the relationships among symptoms within and across clusters may help researchers better understand symptom pathophysiology. Furthermore, if these symptom clusters arise from a common mechanism and are amenable to a common intervention, clinicians can then manage and treat multiple symptoms within the same cluster, which may lead to improvements in quality of life, adherence to treatment, and therapeutic benefit.9-11
In recent years, symptom clusters have been evaluated across a variety of cancers,12-14 with several studies in breast cancer,15-17 lung cancer,18-20 and non-Hodgkin lymphoma21 and across multiple cancer sites.22-32 Each of these studies prospectively collected patient-reported symptoms. Although having patient-reported outcomes is important for the understanding of symptom clusters, they are not always collected in clinical trials. Thus, these trials must be supplemented by additional studies that look specifically at symptom clusters. Moreover, the instruments used to measure patient-reported symptoms varied across the studies. Several studies used the MD Anderson symptom inventory,11,29 whereas others used self-developed symptom checklists.12-26,28 The level of detail with which symptoms are collected and the symptoms that are included vary across these instruments. The lack of uniformity regarding selected symptoms scales and instruments likely contributed to the lack of consistency of the symptom clusters identified.33 Thus, comparisons across studies are challenging, and the studies are hard to reproduce.
In this article, to identify adverse event clusters, we proposed using the adverse event data captured using the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE)34 from large and well-executed clinical trials. The collection of data on adverse events using the CTCAE is standard practice in clinical trials and is another potential way to explore adverse event clusters in the absence of patient-reported outcomes. Aside from the data’s availability, the wide range of adverse event categories reported in the CTCAE allows evaluation of the relationship between symptoms and many clinician-reported adverse events that are not symptoms. To illustrate this, we performed a cluster analysis using the adverse event data from a randomized clinical trial in patients with advanced prostate cancer. Given that adverse events are common and sometimes severe in patients undergoing treatment of advance prostate cancer, the identification of clusters in this population provides useful information for symptom management. Moreover, only one symptom cluster study to date has focused on patients with prostate cancer. This study collected six symptoms from patients with early-stage disease treated with brachytherapy, external beam radiation therapy, or radical prostatectomy.31 In addition, one previous study used the CTCAE to identify symptom clusters in head and neck cancer.32 No research has been performed to identify adverse event clusters in patients with advanced prostate cancer.
METHODS
Patients and Data
This study is a secondary analysis of the data collected in a phase III trial conducted by SWOG-S9916. Six hundred seventy-four patients with prostate cancer who met the eligibility criteria were enrolled onto this study from 1999 to 2003 and were randomly allocated to docetaxel with estramustine (n = 338) or mitoxantrone with prednisone (n = 336). The trial results showed a beneficial effect on survival of docetaxel plus estramustine, making it the new standard of care for pathologically confirmed adenocarcinoma of the prostate and progressive metastatic disease of stage D1 or D2 disease. Details of the trial were published previously.35-37 The adverse event data for the trial were captured using CTCAE version 2.0. The CTCAE is a descriptive terminology that can be used for adverse event grading and reporting. It is organized by system organ class and within each class, adverse events are defined along with severity grades and their respective descriptions. Although each grade is adverse event specific, the CTCAE defines all grades on a scale from 1 to 5: with 1 a mild or asymptomatic adverse event; 2 a moderate adverse event; 3 a severe or medically significant but not life-threatening adverse event; 4 a life-threatening adverse event with consequences; and 5, death. In the trial, adverse event data were captured after each cycle of treatment. However, only the highest-grade adverse event across all cycles for each type of adverse event was entered into the database. Moreover, the data entry of grade 1 and grade 2 adverse events was not required for this trial, leading to unreliable information on low-grade adverse events. Thus, each adverse event category was coded as a binary variable, severe adverse event, defined as the presence or absence of a grade 3 or higher adverse event. Patients without any severe adverse event, or who had incomplete information on the severity of the adverse event and could not be classified, were excluded from the analysis. Thus, patients with grade 1 and grade 2 adverse events only were excluded from the analysis and patients with at least one grade 3 or higher adverse event were included, with each of their adverse events coded as 0 or 1 depending on severity. Although the original trial was conducted more than a decade ago, docetaxel treatment remains a standard of care for patients with advanced prostate cancer today. Thus, the adverse event data remain relevant.
Statistical Analysis
Adverse event data from both arms in the trial were pooled for the analysis. The CTCAE, with its large number of categories, classifies adverse events with a high degree of granularity and thus the prevalence of any individual adverse event is lower than it would be with more generalized categories. Therefore, the individual adverse events captured were reviewed by an oncologist and combined on the basis of clinical similarity; this decreased the number of adverse event categories used in the statistical analysis and increased the prevalence of each category. The combined adverse event categories are displayed in the online-only Appendix Table A1, along with the names of the new collapsed adverse event type. Only adverse event categories with more than 10 events were included in the cluster analysis, given that adverse event categories with less than 10 events may be unstable and unreliable because of small sample size. A higher prevalence threshold was considered for inclusion of adverse event categories but was deemed less useful given that severe adverse events in general are not prevalent, and this may limit the number of adverse events that can be included. A variable-based hierarchical cluster analysis was conducted to combine severe adverse events that tend to occur simultaneously into the same clusters.38 Clusters were formed in such a way that severe adverse events within a cluster were more similar and those from different clusters were more dissimilar. The sum of the correlations among severe adverse events was the metric used to define the extent to which a cluster was evident. A dendrogram was used to summarize the relationships among all the severe adverse event types. The cluster analysis was performed using the R package ClustOfVar.39 The cluster analysis did not control for treatment assignment.
On the basis of the similarity identified in the cluster analysis, Mantel-Haenszel χ2 tests were conducted to evaluate the associations among severe adverse event variables within the same cluster.40 Spearman rank correlations (r) among those severe adverse events within a cluster were also calculated and were expressed as an absolute value between 0 and 1, with 0 representing no correlation and 1 representing perfect correlation.41 After the clusters were identified, Fisher’s exact tests were used to test if the distribution of each cluster differed by treatment. Statistical analyses other than cluster analysis were performed using SAS, version 9.4 (SAS Institute, Cary, NC).
RESULTS
The demographic information for patients enrolled onto this trial was reported previously by Petrylak et al.35 As of December 2003, all surviving patients had stopped the treatment protocol. Three hundred twenty-three patients who had at least one severe adverse event and no missing information on adverse event severity were included in our analysis. The median age of the study cohort was 73 years with a range of 50 to 88 years. The majority of the patients were white (83%) and the rest were African American (15%), Asian (1%), and Native American (< 1%). Eighty-nine percent of these patients had a baseline Eastern Cooperative Oncology Group performance-status score of 0 or 1. The median prostate-specific antigen level at random assignment was 99.2 ng/mL.
One hundred nine distinct adverse events with at least one grade 3 or higher event were captured using the CTCAE grading system. After combining adverse events with clinical similarity and small prevalence, 92 categories of adverse events were derived; these categories, along with their prevalence, are listed in Table 1. Twenty-three of the 92 categories had more than 10 events and thus were included in the cluster analysis. The frequency and percentage of each category are shown in Fig 1. The remaining 69 categories were excluded from the cluster analysis because of low prevalence.
Table 1.
Number and Percentage of Patients With Each Adverse Event Category
Some adverse events categories are defined based on the combination of several adverse event categories reported in National Cancer Institute Common Terminology Criteria for Adverse Events, version 2.0. Details are listed in Appendix Table A1.
Distribution of the adverse events included in the cluster analysis. PRBC, packed RBC; w/o, without.
Severe adverse event clusters with P < .05 from the Mantel-Haenszel χ2 test are shown in the dendogram in Fig 2, along with the correlations among the adverse events. In psychosocial studies, correlations of 0.10 are generally considered weak associations, 0.30 are considered moderate, and 0.50 are considered strong.42 On the basis of this interpretation, four clusters can be considered to have moderate associations. These are (1) nausea, vomiting, and anorexia (n = 35, r = 0.45), 31 (9%) observed in the docetaxel plus estramustine arm (DE) and four (1%) in the mitoxantrone plus prednisone (MP) arm (P < .001); (2) joint/bone and muscle weakness (n = 26, r = 0.29), 19 (6%) in the DE arm and seven (2%) in the MP arm (P = .026); (3) anemia and packed RBC (PRBC) transfusion (n = 20, r = 0.38), 12 (4%) in the DE arm and eight (2%) in the MP arm (P = .50); and (4) neutrophils/granulocytes, febrile neutropenia, and leukocytes/lymphopenia (n = 114, r = 0.29), 59 (17%) in the DE arm and 55 (16%) in the MP arm (P = .76). In addition, there were two additional clusters with weak associations: (1) fatigue/malaise/lethargy and dehydration (n = 66, r = 0.12), 37 (11%) in the DE arm and 29 (9%) in the MP arm (P = .37) and (2) constipation, infection without neutropenia, and abdominal pain/cramping (n = 35, r = 0.13), 22 (7%) in the DE arm and 13 (4%) in the MP arm (P = .16). Dyspnea was also a frequently detected symptom for those who were in the cluster of nausea, vomiting, and anorexia (r = 0.18). Therefore, significantly higher rates of the nausea, vomiting, and anorexia cluster and the joint/bone and muscle weakness cluster were observed in the DE arm compared with the MP arm.
Dendrogram for the cluster analysis. PRBC, packed RBC; w/o, without. Numbers in the figure are Spearman rank correlation, and Ps are from the Mantel-Haenszel χ2 test.
DISCUSSION
We illustrate the feasibility of using adverse event data captured using the CTCAE to identify adverse event clusters in a large phase III clinical trial in advanced prostate cancer. A wide range of adverse events and symptoms is observed in patients with advanced prostate cancer, and little is known about the correlations among them. In addition, patient-reported outcomes from the same trial indicated no statistically significant treatment arm differences for pain palliation and global quality of life.36 However, a subsequent analysis reported that variation in global quality of life remains after controlling for pain and that the relationship between global quality of life and pain differs by treatment.37 Thus, to improve our understanding of these adverse events, the need for more research in this area is clear. Our analysis of clinician-reported severe adverse events in patients with advanced prostate cancer identified several severe adverse event clusters. Some of these clusters have been reported previously in the literature of other types of cancers on the basis of patient-reported symptoms. The cluster of nausea and vomiting is widely reported by multiple studies.11,17 The addition of anorexia to this cluster has also been reported, as has the addition of dyspnea.18,29 The association between fatigue/malaise/lethargy and dehydration has also been reported previously.28 These findings are reassuring and confirm the clinical relevance of these severe adverse event clusters in the management of symptoms. Because of the extent of clinician-reported CTCAE data available for this study, we were able to identify several unique severe adverse event clusters. The high correlation between anemia and PRBC transfusion can be explained based on clinic practice; that is, when patients with prostate cancer are anemic, they are more likely to get transfusions. Thus, transfusions may be considered more as interventions rather than as simultaneous adverse events. On the other hand, the cluster of neutrophils/granulocytes, leukocytes/lymphopenia, and febrile neutropenia is consistent with bone marrow suppression from chemotherapy, which increases the risk of neutropenic fever. A sensitivity analysis was conducted by collapsing anemia with PRBC transfusion and all leukopenic events. Collapsing these categories did not affect the remaining clusters. The cluster formed by joint/bone and muscle weakness is informative for symptom management because it implies that patients who experience joint and muscle pain may also experience muscle weakness. The associations among constipation, abdominal pain/cramping, and infection without neutropenia may be explained by the fact that a large proportion of patients who had infection without neutropenia also had a urinary tract infection. Given the weak association of the last cluster, it should be further investigated in other settings.
The Maliski et al31 study in prostate cancer on 406 patients with early-stage disease only collected six symptoms, which had been grouped already to some extent. These symptoms were emotional distress, fatigue, pain, urinary dysfunction, sexual dysfunction, and bowel dysfunction. In our study, the wide array of adverse events allowed us to find new adverse event clusters. For example, whereas Maliski et al31 had a symptom called bowel dysfunction, we had more granular data with adverse events such as constipation, abdominal pain/cramping, and diarrhea. Having more detailed adverse events information allowed us to explore more varied associations, which can be overlooked with combined symptom categories. The large number of adverse event types included in the CTCAE can also be challenging in cluster analysis because of the sparseness of the data. Thus, it becomes necessary to combine categories with clinical similarity to increase the prevalence of adverse events with low prevalence. The way in which adverse events are combined can affect the resulting clusters. Therefore, when categories are combined in a subjective manner, it may be necessary to explore alternative ways of collapsing adverse event types. In this study, we conducted an alternative analysis by reclassifying infection by sites (urinary, respiratory, and other infection) instead of with/without neutropenia, but no additional cluster was identified. Given the subjectivity of the infection categories, the clusters that include them should be interpreted with caution. Another difficulty of using adverse event data from clinical trials, which is also a limitation of our study, is that often not all adverse events are captured in the database. For example, in our study, only the highest severity grade is recorded for each adverse event type and only grade 3 or higher adverse events are reliably captured. Thus, the adverse events reported by the same individual may have occurred at different time points, and there was a lack of information regarding low-grade adverse events. Further research is needed to evaluate adverse event clusters in a prospective longitudinal manner and to assess the relationships among these adverse event clusters, clinical outcomes, the change over time in adverse event status, and the impact of low-grade adverse events. Another interesting topic is whether the severe adverse event cluster is associated with clinical outcomes for patients with advanced prostate cancer. To answer this question, a time-to-event analysis should be performed to better understand the role of adverse event clusters as a prognostic factor. Eton et al27 reported that early deterioration in symptom status was prognostic for lung cancer clinical outcomes. Given the lack of information on the time of the highest-grade adverse event experienced, we were not able to clearly explore this point in this study.
Although the CTCAE allows us to evaluate a wide range of adverse events, it is clinician reported and may not accurately reflect the presence of patient symptoms and the severity of these symptoms. Thus, it is possible that patient-reported symptoms were disregarded and not reported by clinicians, and symptoms that were reported as severe by patients were deemed not severe by clinicians. The discrepancy between patient-reported symptoms and clinician-reported symptoms is a limitation of this study. Moreover, the severe adverse event clusters identified in this study are based on grading descriptions provided in the CTCAE version 2.0 and only consider grade 3 or higher adverse events. Therefore, the interpretation of these results beyond the CTCAE 2.0 and severe adverse events should be performed with caution because they may not apply directly, and thus, different clusters may be identified. Moreover, adverse event categories with sparse data and a smaller number of events are less reliable and more variable. Thus, clusters that include them should be confirmed in future studies. Replication and extension of this work using data from a recent prostate cancer trial are encouraged to validate the clusters identified. Another avenue of research would be to compare clusters identified with the CTCAE versus the Patient-Reported Outcomes–CTCAE items developed by the National Cancer Institute.43
In conclusion, we identified several severe adverse event clusters in patients with advanced prostate cancer and demonstrated the feasibility of using the CTCAE data from a large clinical trial to identify adverse event clusters. Importantly, these clusters cross the system organ class of the CTCAE, which is based on an anatomic or physiologic system and may not imply common symptom pathophysiology. Therefore, the clusters expand our understanding of the association among adverse events and may indicate that adverse events within each cluster should be treated and managed together.
Acknowledgment
Supported in part by the American Cancer Society (Grant No. MRSG-13-146-01-CPHPS) and in part by the National Cancer Institute Community Oncology Research Program awarded to SWOG (Grant No. 1UG1CA189974-01).
Appendix
Table A1.
Combined Adverse Event Categories
Collapsed Adverse Event Category/Adverse Event Category in CTCAE
Includes only infections with grade 3 or 4 neutropenia.
AUTHOR CONTRIBUTIONS
Conceptionand design: Xiaobo Zhong, Dawn L. Hershman, Joseph Unger, Shing M. Lee
Financial support: Shing M. Lee
Collection and assembly of data: Xiaobo Zhong
Data analysis and interpretation: Xiaobo Zhong, Emerson A. Lim, Carol M. Moinpour, Joseph Unger, Shing M. Lee
Manuscript writing: All authors
Final approval of manuscript: All authors
AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
Identifying Severe Adverse Event Clusters Using the National Cancer Institute's Common Terminology Criteria for Adverse Events
The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or jop.ascopubs.org/site/misc/ifc.xhtml.
Xiaobo Zhong
No relationship to disclose
Emerson A. Lim
Honoraria: Dendreon
Consulting or Advisory Role: Dendreon
Dawn L. Hershman
No relationship to disclose
Carol M. Moinpour
Stock or Other Ownership: Affymetrix, Amgen, Dendreon, Merck, Nanogen, PDL BioPharma, Pfizer (I)
Joseph Unger
No relationship to disclose
Shing M. Lee
No relationship to disclose
References
1.Lebret T Culine S Davin JL, etal: Quality of life of 1276 elderly patients with prostate cancer, starting treatment with a gonadotropin-releasing hormone agonist: Results of a French observational study
Aging Male
17:87–93,2014 [DOI] [PubMed] [Google Scholar]
2.Litwin MS Sadetsky N Pasta DJ, etal: Bowel function and bother after treatment for early stage prostate cancer: A longitudinal quality of life analysis from CaPSURE
J Urol
172:515–519,2004 [DOI] [PubMed] [Google Scholar]
3.Galbraith ME, Ramirez JM, Pedro LW: Quality of life, health outcomes, and identity for patients with prostate cancer in five different treatment groups
Oncol Nurs Forum
28:551–560,2001 [PubMed] [Google Scholar]
4.Petry H Berry DL Spichiger E, etal: Responses and experiences after radical prostatectomy: Perceptions of married couples in Switzerland
Int J Nurs Stud
41:507–513,2004 [DOI] [PubMed] [Google Scholar]
5.Weber BA Roberts BL Resnick M, etal: The effect of dyadic intervention on self-efficacy, social support, and depression for men with prostate cancer
Psychooncology
13:47–60,2004 [DOI] [PubMed] [Google Scholar]
6.Nelson WG, De Marzo AM, Isaacs WB: Prostate cancer
N Engl J Med
349:366–381,2003 [DOI] [PubMed] [Google Scholar]
7.Esper P: Symptom clusters in individuals living with advanced cancer
Semin Oncol Nurs
26:168–174,2010 [DOI] [PubMed] [Google Scholar]
8.Xiao C: The state of science in the study of cancer symptom clusters
Eur J Oncol Nurs
14:417–434,2010 [DOI] [PubMed] [Google Scholar]
9.Aktas A: Cancer symptom clusters: Current concepts and controversies
Curr Opin Support Palliat Care
7:38–44,2013 [DOI] [PubMed] [Google Scholar]
10.Miaskowski C, Dodd M, Lee K: Symptom clusters: The new frontier in symptom management research
J Natl Cancer Inst Monogr
2004:17–21,2004 [DOI] [PubMed] [Google Scholar]
11.Chen ML, Tseng HC: Symptom clusters in cancer patients
Support Care Cancer
14:825–830,2006 [DOI] [PubMed] [Google Scholar]
12.Miller DC Hafez KS Stewart A, etal: Prostate carcinoma presentation, diagnosis, and staging: An update form the National Cancer Data Base
Cancer
98:1169–1178,2003 [DOI] [PubMed] [Google Scholar]
13.Teunissen SCCM de Graeff A Voest EE, etal: Are anxiety and depressed mood related to physical symptom burden? A study in hospitalized advanced cancer patients
Palliat Med
21:341–346,2007 [DOI] [PubMed] [Google Scholar]
14.Davis MP, Kirkova J: Lifting symptom burden--how far off the ground are we?
Support Care Cancer
16:757–761,2008 [DOI] [PubMed] [Google Scholar]
15.Doong SH Dhruva A Dunn LB, etal: Associations between cytokine genes and a symptom cluster of pain, fatigue, sleep disturbance, and depression in patients prior to breast cancer surgery
Biol Res Nurs
17:237–247,2015 [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Bender CM Ergÿn FS Rosenzweig MQ, etal: Symptom clusters in breast cancer across 3 phases of the disease
Cancer Nurs
28:219–225,2005 [DOI] [PubMed] [Google Scholar]
17.Wilmoth MC Coleman EA Smith SC, etal: Fatigue, weight gain, and altered sexuality in patients with breast cancer: Exploration of a symptom cluster
Oncol Nurs Forum
31:1069–1075,2004 [DOI] [PubMed] [Google Scholar]
18.Gift AG Stommel M Jablonski A, etal: A cluster of symptoms over time in patients with lung cancer
Nurs Res
52:393–400,2003 [DOI] [PubMed] [Google Scholar]
19.Gift AG Jablonski A Stommel M, etal: Symptom clusters in elderly patients with lung cancer
Oncol Nurs Forum
31:202–212,2004 [DOI] [PubMed] [Google Scholar]
20.Maguire R Stoddart K Flowers P, etal: An interpretative phenomenological analysis of the lived experience of multiple concurrent symptoms in patients with lung cancer: A contribution to the study of symptom clusters
Eur J Oncol Nurs
18:310–315,2014 [DOI] [PubMed] [Google Scholar]
21.Bolukbas F, Kutluturkan S: Symptoms and symptom clusters in non Hodgkin’s lymphoma patients in Turkey
Asian Pac J Cancer Prev
15:7153–7158,2014 [DOI] [PubMed] [Google Scholar]
22.Dong ST Butow PN Costa DS, etal: Symptom clusters in patients with advanced cancer: A systematic review of observational studies
J Pain Symptom Manage
48:411–450,2014 [DOI] [PubMed] [Google Scholar]
23.Alesi ER, del Fabbro E: Opportunities for targeting the fatigue-anorexia-cachexia symptom cluster
Cancer J
20:325–329,2014 [DOI] [PubMed] [Google Scholar]
24.Fishbain D Gao JR Lewis J, etal: Examination of symptom clusters in acute and chronic pain patients
Pain Physician
17:E349–E357,2014 [PubMed] [Google Scholar]
25.Olver IN, Eliott JA, Koczwara B: A qualitative study investigating chemotherapy-induced nausea as a symptom cluster
Support Care Cancer
22:2749–2756,2014 [DOI] [PubMed] [Google Scholar]
26.Kim HJ, Malone PS, Barsevick AM: Subgroups of cancer patients with unique pain and fatigue experiences during chemotherapy
J Pain Symptom Manage
48:558–568,2014 [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Eton DT Fairclough DL Cella D, etalEastern Cooperative Oncology Group : Early change in patient-reported health during lung cancer chemotherapy predicts clinical outcomes beyond those predicted by baseline report: Results from Eastern Cooperative Oncology Group Study 5592
J Clin Oncol
21:1536–1543,2003 [DOI] [PubMed] [Google Scholar]
28.Aktas A, Walsh D, Hu B: Cancer symptom clusters: an exploratory analysis of eight statistical techniques
J Pain Symptom Manage
48:1254–1266,2014 [DOI] [PubMed] [Google Scholar]
29.Ivanova MO Ionova TI Kalyadina SA, etal: Cancer-related symptom assessment in Russia: Validation and utility of the Russian M. D. Anderson Symptom Inventory
J Pain Symptom Manage
30:443–453,2005 [DOI] [PubMed] [Google Scholar]
30.Walsh D, Rybicki L: Symptom clustering in advanced cancer
Support Care Cancer
14:831–836,2006 [DOI] [PubMed] [Google Scholar]
31.Maliski SL Kwan L Elashoff D, etal: Symptom clusters related to treatment for prostate cancer
Oncol Nurs Forum
35:786–793,2008 [DOI] [PubMed] [Google Scholar]
32.Xiao C Hanlon A Zhang Q, etal: Symptom clusters in patients with head and neck cancer receiving concurrent chemoradiotherapy
Oral Oncol
49:360–366,2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Fan G, Filipczak L, Chow E: Symptom clusters in cancer patients: A review of the literature
Curr Oncol
14:173–179,2007 [DOI] [PMC free article] [PubMed] [Google Scholar]
34.National Cancer Institute : CTEP. Common Terminology Criteria for Adverse Events version 2.0
1998.
Bethesda, MD:
National Institutes of Health [Google Scholar]
35.Petrylak DP Tangen CM Hussain MH, etal: Docetaxel and estramustine compared with mitoxantrone and prednisone for advanced refractory prostate cancer
N Engl J Med
351:1513–1520,2004 [DOI] [PubMed] [Google Scholar]
36.Berry DL Moinpour CM Jiang CS, etalSouthwest Oncology Group : Quality of life and pain in advanced stage prostate cancer: Results of a Southwest Oncology Group randomized trial comparing docetaxel and estramustine to mitoxantrone and prednisone
J Clin Oncol
24:2828–2835,2006 [DOI] [PubMed] [Google Scholar]
37.Moinpour CM, Donaldson GW, Redman MW: Do general dimensions of quality of life add clinical value to symptom data?
J Natl Cancer Inst Monogr
2007:31–38,2007 [DOI] [PubMed] [Google Scholar]
39.Chavent M Kuentz V Liquet B, etal: ClustOfVar: An R package for the clustering of variables
J Stat Softw
50:1–16,2012. 25317082 [Google Scholar]
40.Wallenstein S, Wittes J: The power of the Mantel-Haenszel test for grouped failure time data
Biometrics
49:1077–1087,1993 [PubMed] [Google Scholar]
41.Spearman C: The proof and measurement of association between two things
By C. Spearman, 1904. Am J Psychol 100:441-471, 1987 [PubMed] [Google Scholar]
42.Cohen J.: Statistical Power Analysis for the Behavioral Sciences
1988.
ed 2
Hillsdale, NJ:
Lawrence Erlbaum [Google Scholar]
43.Basch E Reeve BB Mitchell SA, etal: Development of the National Cancer Institute’s patient-reported outcomes version of the common terminology criteria for adverse events (PRO-CTCAE)
J Natl Cancer Inst
2014106:dju244 [DOI] [PMC free article] [PubMed] [Google Scholar]
